News for hackers. Stuff that matters.
Re:Low-code, no-model recommendation system with GPT3 by superasn on Wednesday September 09, 2020 @06:22AM (#24419460)
Thanks for the link. So as someone who is completely new to this, if this guy used this model instead of gpt 3 would he get similar results?
Re:Low-code, no-model recommendation system with GPT3 by calebkaiser on Wednesday September 09, 2020 @10:48AM (#24421612)
If you want to play with your own GPT-2 API, I put this together and shared it on HN yesterday: https://github.com/cortexlabs/cortex/tree/master/examples/py...

You can run that and deploy your own GPT-2 API locally, or in the cloud. I'd also recommend checking out Max Woolf's gpt-2-simple, which is a library that makes it really easy to fine tune GPT-2 with your own text (if you've played the free version of AI Dungeon, this is how they trained their model):

https://github.com/minimaxir/gpt-2-simple

Re:Low-code, no-model recommendation system with GPT3 by dhairya on Wednesday September 09, 2020 @06:29AM (#24419506)
out of the box probably not. gpt-3 is pretrained on a much larger corpus and is a much higher capacity model, which is why its been performing very well in zero-shot learning examples like the one above. That being said you should try it.

Note the huggingface library is a python library, so you'd technically be writing some code, though the library API is super user friendly and the community is very helpful.

Re:Low-code, no-model recommendation system with GPT3 by sanxiyn on Wednesday September 09, 2020 @06:28AM (#24419503)
No, because the model is smaller.